[Illustration: Fig. 35. How Cell Walls Shrink.] (2) Thick-walled cells shrink more than thin-walled cells, that is, summer cells more than spring cells. This is due to the fact that they contain more shrinkable substance. The thicker the wall, the more the shrinkage. Consider the effects of these changes; ordinarily a log when drying begins to "check" at the end. This is to be explained thus: Inasmuch as evaporation takes place faster from a cross than from a longitudinal section, because at the cross-section all the cells are cut open, it is to be expected that the end of a piece of timber, Fig. 36, A, will shrink first. This would tend to make the end fibers bend toward the center of the piece as in B, Fig. 36. But the fibers are stiff and resist this bending with the result that the end splits or "checks" as in C, Fig. 36. But later, as the rest of the timber dries out and shrinks, it becomes of equal thickness again and the "checks" tend to close. [Illustration: Fig. 36. The Shrinkage and Checking at the End of a Beam.] (3) For some reason, which has not been discovered, the cells or fibers of wood do not shrink in length to any appreciable extent. This is as true of the cells of pith rays, which run radially in the log, as of the ordinary cells, which run longitudinally in it. In addition to "checking" at the end, logs ordinarily show the effect of shrinkage by splitting open radially, as in Fig. 37. This is to be explained by two factors, (1) the disposition of the pith (or medullary) rays, and (2) the arrangement of the wood in annual rings. [Illustration: Fig. 37. The Shrinkage and Splitting of a Log.] (1) The cells of the pith rays, as we have seen in Chapter I, run at right angles to the direction of the mass of wood fibers, and since they shrink according to the same laws that other cells do, viz., by the cell wall becoming thinner but not shorter, the strain of their shrinkage is contrary to that of the main cells. The pith rays, which consist of a number of cells one above the other, tend to shrink parallel to the length of the wood, and whatever little longitudinal shrinkage there is in a board is probably due mostly to the shrinkage of the pith rays. But because the cells of pith rays do not appreciably shrink in their length, this fact tends to prevent the main body of wood from shrinking radially, and the result is that wood shrinks less radially than tangentially. Tangentially is the only way left for it to shrink. The pith rays may be compared to the ribs of a folding fan, which keep the radius of unaltered length while permitting comparative freedom for circumferential contraction.